On Fri, Jun 04, 2010 at 03:06:30PM -0700, Gary Greene wrote:
On 6/4/10 2:59 PM, "m.roth@5-cent.us" m.roth@5-cent.us wrote:
Thanks, all, and with the help of the other admin, the system is up. What I had to do was linux rescue, the chroot /mnt/sysimage, grub-install /dev/sda
What I didn't get until later was it also needed /boot/grub/grub.conf, and then ln -s /boot/grub/grub.conf /boot/grub/menu.lst ln -s /boot/grub/grub.conf /etc/grub.conf
and it now boots. As I said, *so* much easier than editing /etc/lilo.conf, and rerunning lilo....
mark
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Mark, you might dislike Grub, but overall, it's saved me a lot of grief since lilo cannot start a "shell" so the poor sod can coerce the system to start another kernel if needed. (Yes, this has happened more than on one occasion with many the distro for me when dealing with rolling my own kernel.) That and lilo is completely useless on newer hardware (EFI and GPT labels anyone?) since it only understands BIOS addresses, whereas GRUB2 understands both. As stated in an earlier email in this thread, this is mostly caused by Anaconda and some of the GUI tools doing the wrong thing, not GRUB.
Lilo is good choice once you're running software raid. It can write MBR on all disks and kernel can boot only from one disk with no intervation. Grub's writing MBR only at the first disk.